CN-121980699-A - Water turbine operation evaluation method and system based on multi-source data fusion

CN121980699ACN 121980699 ACN121980699 ACN 121980699ACN-121980699-A

Abstract

The invention discloses a water turbine operation evaluation method and system based on multi-source data fusion, and relates to the technical field of water turbine evaluation. A water turbine operation evaluation system based on multi-source data fusion comprises a data acquisition module, a physical constraint module, a pressure wave hysteresis compensation module, a sediment degradation module, a characteristic relation module and an evaluation result module. According to the invention, under the condition of correcting the water head and efficiency after hysteresis compensation and sediment compensation, the ideal characteristic relation of the opening degree and the active power of the guide vane is fitted based on stable operation history data, the actual operation track is compared with the ideal characteristic curve in a deviation manner, and the abnormal condition is identified by utilizing the amplitude, the sign and the time sequence characteristic of the deviation, so that the on-line fine diagnosis of the states of the guide vane and the servo system is realized.

Inventors

LIU XING
ZHU JIAYING
SHE WEIWEI
SONG RUIQI
DONG XIN
HE JUN

Assignees

中国水利电力对外有限公司
中水电电力发展有限公司

Dates

Publication Date: 20260505
Application Date: 20260108

Claims (10)

1. A water turbine operation evaluation method based on multi-source data fusion is characterized by constructing a twin body of a water turbine, representing the operation state of the twin body by a state vector, continuously correcting and updating the state vector, and comprising the following steps: The method comprises the steps of collecting observation data related to water head, flow, unit output, guide vane opening instruction and feedback, vibration, servo oil pressure, electric parameters and sediment, synchronously calibrating the observation data, and distributing credibility as the input of a twin body; Constructing a physical constraint model based on hydraulic power and mechanical mechanism of the water turbine, carrying out physical consistency correction and fusion updating on the state vector based on the observation data and the corresponding credibility, and generating a physical residual error; According to the structure of the water diversion system, the water body characteristics and the physical residual errors, determining the propagation delay of the water head, performing hysteresis compensation on components related to the water head in the state vector, and adaptively adjusting the pressure wave model parameters of the twin body; building a sediment degradation model, compensating components related to efficiency and vibration in a state vector, predicting performance degradation trend, and feeding back vibration change caused by sediment to a physical constraint model; Based on the ideal characteristic relation, carrying out deviation comparison, identifying abnormality of the guide vane or a servo system, and updating guide vane and servo related parameters of the state vector; based on the update of the state vector, the operation evaluation result, the abnormal alarm, the historical state backtracking and the performance prediction information are output.
2. The method for evaluating the operation of the water turbine based on the multi-source data fusion is characterized in that the synchronous calibration of the observation data comprises the steps of establishing a unified time reference by carrying out offset correction and interpolation compensation on original timestamps of different data sources, realizing sampling frequency uniformity on different observation data according to the unified time reference in a mode of up-sampling, down-sampling or sliding window interpolation, and carrying out event alignment processing on time sequences of different types of observation data by using operation events of guide vane opening mutation, unit output change, water head jump or load dump as synchronous trigger points.
3. The method for evaluating the operation of the water turbine based on the multi-source data fusion according to claim 1, wherein the allocation of the reliability comprises calculating initial reliability for each type of observed data according to noise level, signal fluctuation degree, long-term stability and historical reliability of a corresponding sensor of each type of observed data, and dynamically adjusting the initial reliability when short-time interruption, amplitude abnormal fluctuation or significant inconsistency exists between the observed data and adjacent data sources.
4. The method for evaluating the operation of the water turbine based on multi-source data fusion according to claim 1 is characterized in that the construction of the physical constraint model comprises the steps of determining physical correlation variables among a water head, a flow, a unit output, a guide vane opening, a servo oil pressure and an electric parameter according to design parameters and operation characteristics of the water turbine, establishing a constraint equation describing relations among the water head, the flow, the unit output, the guide vane opening, the runner hydraulic power action and the unit output and the electric parameter based on hydraulics and a mechanical transmission mechanism, and identifying and correcting parameters in the constraint equation by utilizing historical operation data or test calibration data of the water turbine to form the physical constraint model for carrying out physical consistency correction on a state vector.
5. The method for evaluating the operation of the water turbine based on the multi-source data fusion according to claim 1, wherein the generating of the physical residual errors comprises the steps of calculating physical pre-measurement corresponding to a water head, a flow rate, a unit output, a guide vane opening, a servo oil pressure and an electric parameter based on a current state vector by using a physical constraint model, comparing the physical pre-measurement with corresponding observed data subjected to synchronous calibration, obtaining deviation values of the physical quantities according to a predetermined rule, and carrying out normalization processing and classification sorting on the deviation values according to different physical links of hydraulic power, mechanical power and electric power conversion to form a physical residual error vector indicating the consistency degree of the physical links.
6. The method for evaluating the operation of the water turbine based on multi-source data fusion according to claim 1 is characterized in that hysteresis compensation is carried out by determining equivalent hydraulic path length from a water head measuring point to the water turbine according to a pre-acquired water diversion system structural parameter, acquiring initial pressure wave propagation characteristics by combining calibrated water body characteristic parameters, correcting the initial pressure wave propagation characteristics by utilizing phase deviation or time deviation characteristics related to a water head in a physical residual error, estimating propagation delay of water head data, carrying out time axis translation or resampling processing on the synchronously calibrated water head data according to the estimated propagation delay, writing the corrected water head data into components related to the water head in a state vector, completing hysteresis compensation on the water head components, and continuously counting corresponding relations between physical residual errors and propagation delay estimation under a plurality of operation conditions, and carrying out self-adaptive adjustment on wave speed and attenuation coefficients of a pressure wave model in a twin body.
7. The method for evaluating the operation of the water turbine based on the multi-source data fusion according to claim 1, wherein the compensation of the components related to the efficiency and the vibration in the state vector and the prediction of the performance degradation trend comprise the steps of establishing a sediment degradation model representing the influence relation of sediment concentration, particle characteristics and flow velocity on the efficiency decline and the vibration amplitude change of the unit based on observed data related to sediment and efficiency change and vibration response in a historical operation process, calculating the efficiency loss amount and the additional vibration component caused by sediment under the current working condition by using the sediment degradation model, deducting or correcting the efficiency loss amount and the additional vibration component from the corresponding efficiency component and vibration component in the state vector to obtain the efficiency characteristic and the vibration characteristic after the compensation of the sediment influence, and predicting the efficiency degradation trend and the vibration enhancement trend of the subsequent operation stage by extrapolating the sediment degradation model over a plurality of operation cycles according to the accumulation effect of sediment working conditions, and using the vibration characteristic change caused by sediment to correct the parameters related to the vibration response in the physical constraint model.
8. The method for evaluating the operation of the water turbine based on the multi-source data fusion according to claim 1, wherein the generation of the ideal characteristic relation of the opening degree and the power of the guide vane comprises the steps of selecting a water head component, a flow component and an efficiency component which are subjected to hysteresis compensation and sediment degradation compensation and a guide vane opening degree component and an active power component which correspond to the water head component, the flow component and the efficiency component from a state vector, removing operation condition data with abnormal fluctuation, fitting or interpolating the corresponding relation between the opening degree and the active power of the guide vane based on historical operation data which meet the stable operation condition of the water turbine, and constructing an ideal characteristic curve or a function expression form which monotonically changes between the opening degree and the active power of the guide vane under the given correction water head and efficiency conditions as the ideal characteristic relation of the opening degree and the power of the guide vane.
9. The method for evaluating the operation of the water turbine based on the multi-source data fusion according to claim 1 is characterized in that the deviation comparison based on the ideal characteristic relation comprises the steps of determining an operation track of an actual guide vane opening according to guide vane opening instructions and guide vane opening feedback, calculating the deviation of an ideal characteristic relation between active power and guide vane opening-power under the corresponding working condition of the operation track, extracting the amplitude, the sign, the duration and the distribution characteristics changing along with the working condition of the deviation, judging the deviation as guide vane clamping stagnation abnormality when the deviation is saturated in a section mode in the guide vane opening direction or the actual guide vane opening is basically unchanged when the guide vane opening instructions change, judging the deviation as servo system hysteresis abnormality when the deviation is mainly represented as a obvious hysteresis area or a hysteresis loop is formed relative to the guide vane opening instruction signals, and correcting parameters related to guide vane opening dynamics and servo response in a state vector according to the judged abnormal type.
10. A system for evaluating the operation of a water turbine based on multi-source data fusion, wherein the system is applied to the method for evaluating the operation of a water turbine based on multi-source data fusion according to any one of claims 1 to 9, and the system comprises: The data acquisition module acquires observation data related to water head, flow velocity, guide vane opening instruction and feedback, vibration, oil pressure, electric parameters and sediment, synchronously calibrates the observation data, and distributes credibility as the input of a twin body; the physical constraint module is used for constructing a physical constraint model based on hydraulic and mechanical mechanisms of the water turbine, carrying out physical consistency correction and fusion updating on the state vector based on the observation data and the corresponding credibility, and generating a physical residual error; The pressure wave hysteresis compensation module is used for determining the propagation delay of the water head according to the pre-acquired diversion system structure, the calibrated water body characteristics and the physical residual error, performing hysteresis compensation on components related to the water head in the state vector, and adaptively adjusting pressure wave model parameters of the twin body; the sediment degradation module is used for constructing a sediment degradation model, compensating components related to efficiency and vibration in a state vector, predicting performance degradation trend, and feeding back vibration change caused by sediment to the physical constraint model; the characteristic relation module is used for generating an ideal characteristic relation of the opening degree and the power of the guide vane based on the state vector, carrying out deviation comparison based on the ideal characteristic relation, identifying the abnormality of the guide vane or a servo system, and updating the guide vane and the servo related parameters of the state vector; And the evaluation result module is used for outputting operation evaluation results, abnormal alarms, historical state backtracking and performance prediction information based on the update of the state vector.

Description

Water turbine operation evaluation method and system based on multi-source data fusion Technical Field The invention relates to the technical field of water turbine evaluation, in particular to a water turbine operation evaluation method and system based on multi-source data fusion. Background The water turbine is used as core equipment of the hydropower station, and the operation condition of the water turbine is directly related to the safety and stability of the unit and the economy of the power station. In order to improve the operation management refinement level, part of engineering begins to try to analyze and evaluate the state of a unit by using a mechanism model or a data driving model, but the single-point monitoring, the experience judgment and the dispersion index evaluation are still mainly used as the whole, and the unified quantitative description of the operation state of the water turbine is difficult to form. Under the conditions of complex hydraulic structures and variable working conditions, the existing evaluation method has differences in time stamps, sampling frequencies and measurement precision for observed data from different sources, the noise level, reliability and long-term stability of a sensor are different, unified processing is difficult to achieve under the same time scale and reliability framework, asynchronous or contradictory situations often occur among multi-source data, and the existing evaluation method is mostly based on statistical features or local experience curves only and lacks a mechanism for integrating a plurality of variables into a unified physical constraint model to conduct consistency check and state correction. For propagation delay of a water head signal in a long water diversion system, characteristic change of pressure waves and accumulated degradation effect of sediment on unit efficiency reduction and vibration amplitude increase, the existing method lacks explicit modeling, so that performance change caused by sediment degradation, control system abnormality or working condition fluctuation is difficult to distinguish, and the accuracy and the interpretability of an operation evaluation result are limited. Disclosure of Invention The invention provides a hydraulic turbine operation evaluation method which is used for fusing multisource observation data under the unified time reference and credibility weight on the premise of fully considering the long-term influence of water head propagation delay, sediment working condition on efficiency and vibration of a water diversion system and the dynamic characteristics of a guide vane/servo system, constructing a physical constraint model by depending on hydraulic power and mechanical mechanism of a hydraulic turbine, and continuously correcting and updating the state vector representing the state of a unit; in order to achieve the above purpose, the technical scheme adopted by the invention is as follows: A water turbine operation evaluation method based on multi-source data fusion constructs a twin body of a water turbine, characterizes the operation state of the twin body by a state vector, continuously corrects and updates the state vector, and comprises the following steps: The method comprises the steps of collecting observation data related to water head, flow, unit output, guide vane opening instruction and feedback, vibration, servo oil pressure, electric parameters and sediment, synchronously calibrating the observation data, and distributing credibility as the input of a twin body; Constructing a physical constraint model based on hydraulic power and mechanical mechanism of the water turbine, carrying out physical consistency correction and fusion updating on the state vector based on the observation data and the corresponding credibility, and generating a physical residual error; According to the structure of the water diversion system, the water body characteristics and the physical residual errors, determining the propagation delay of the water head, performing hysteresis compensation on components related to the water head in the state vector, and adaptively adjusting the pressure wave model parameters of the twin body; building a sediment degradation model, compensating components related to efficiency and vibration in a state vector, predicting performance degradation trend, and feeding back vibration change caused by sediment to a physical constraint model; Based on the ideal characteristic relation, carrying out deviation comparison, identifying abnormality of the guide vane or a servo system, and updating guide vane and servo related parameters of the state vector; based on the update of the state vector, the operation evaluation result, the abnormal alarm, the historical state backtracking and the performance prediction information are output. The method comprises the steps of establishing a unified time reference by carrying out offset correction and interpolation compensation on o